1. Communication
Macromolecular
Rapid Communications
wileyonlinelibrary.com 1689© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim DOI: 10.1002/marc.201600440
studied polymer system fabricated to date with this block
copolymer self-assembly and nonsolvent induced phase
separation (SNIPS) process[2]
is the diblock copolymer,
poly(styrene-b-(4-vinyl)pyridine) (PS-b-P4VP, SV). The
phase inversion technique was successfully applied to
other diblock copolymer systems including poly(styrene-
b-(dimethylaminoethyl) methacrylate) (PS-b-PDMAEMA),[3]
poly(styrene-b-ethylene oxide) (PS-b-PEO),[4]
poly(styrene-
b-methyl methacrylate) (PS-b-PMMA),[5]
as well as triblock
terpolymer systems such as poly(isoprene-b-styrene-
b-4-vinylpyridine) (PI-b-PS-b-P4VP, ISV),[6]
poly(isoprene-
b-styrene-b-N,N-dimethylacrylamide) (PI-b-PS-b-PDMA)[7]
and poly(styrene-b-4-vinylpyridine-b-propylene sulfide)
(PS-b-P4VP-b-PPS, SVPS).[8]
Employing a second component in the casting dope
solution, various membrane properties were tailored. For
example, the morphology of the top surface layer was
tuned by the addition of small organic molecules[9] and
Deviating from the traditional formation of block copolymer derived isoporous membranes
from one block copolymer chemistry, here asymmetric membranes with isoporous surface
structure are derived from two chemically distinct block copolymers blended during standard
membrane fabrication. As a first proof of principle, the fabrication of asymmetric membranes is
reported, which are blended from two chemically distinct triblock terpolymers, poly(isoprene-
b-styrene-b-(4-vinyl)pyridine) (ISV) and poly(isoprene-b-styrene-b-(dimethylamino)ethyl
methacrylate) (ISA), differing in the pH-responsive hydrophilic segment. Using block
copolymer self-assembly and nonsolvent induced phase separation process, pure and blended
membranes are prepared by varying weight ratios of ISV to ISA. Pure and blended mem-
branes exhibit a thin, selective layer of pores above a macroporous substructure. Observed
permeabilities at varying pH values of blended membranes depend on relative triblock ter-
polymer composition. These results open a new direction for membrane fabrication through
the use of mixtures of chemi-
cally distinct block copolymers
enabling the tailoring of mem-
brane surface chemistries and
functionalities.
Asymmetric Membranes from Two Chemically
Distinct Triblock Terpolymers Blended during
Standard Membrane Fabrication
Yuk Mun Li, Divya Srinivasan, Parth Vaidya, Yibei Gu, Ulrich Wiesner*
Y. M. Li
Robert Frederick Smith School of Chemical
and Biomolecular Engineering
Cornell University
Ithaca, NY 14853, USA
D. Srinivasan, P. Vaidya, Dr. Y. Gu, Prof. U. Wiesner
Department of Materials Science and Engineering
Cornell University
Ithaca, NY 14853, USA
E-mail: ubw1@cornell.edu
1. Introduction
There has been growing interest in asymmetric ultrafiltra-
tion (UF) membranes derived from block copolymers (BCPs)
since the first report by Peinemann et al.[1]
resulting in
isoporous membranes that combine a highly ordered sur-
face structure with high permselectivity. The most widely
Macromol. Rapid Commun. 2016, 37, 1689−1693

2. Y. M. Li et al.
Macromolecular
Rapid Communications
www.mrc-journal.de
www.MaterialsViews.com1690 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
metal ions which form metal-polymer complexes.[10] With
the introduction of an additive that chemically interacts/
swells one block of the block copolymer, pore sizes were
tailored.[11] For example, an organic additive, glycerol,
was added to the ISV system in order to tailor the pore
size moving the range of filtration from UF to nanofiltra-
tion.[12] Pore sizes were also tailored in ISV terpolymer
derived membranes by blending a homopolymer[6] or in
SV copolymer derived membranes by blending of other
SV block copolymers with varied molar mass and block
volume fractions.[13] Finally, organic–inorganic hybrid
membranes have been reported from the SNIPS process
by mixing in sol or other inorganic nanoparticles into
the block copolymer dough,[14] while carbon materials
with asymmetric structure were derived from adding
in resols[15] and subsequent heat processing at elevated
temperatures.
To the best of our knowledge, however, to date no
studies have been reported in which the SNIPS process
was applied to the mixture of two or more chemically
distinct block copolymers. This would be particularly
interesting for mixtures in which the blocks that end up
decorating the pore walls would be chemically distinct. In
this case it would be possible to mix and match different
chemistries and therefore different functionalities and
properties to the pore walls during standard membrane
fabrication. Previously, systems that desired, e.g., a com-
bination of chemistries and/or stimuli responsive perfor-
mances in the pore walls were limited to an extra post-
modification grafting step on the final membrane.[8,16]
In
contrast, here we demonstrate the facile approach for the
fabrication of membranes that exhibit two chemistries in
the pores as a consequence of incorporating two chemi-
cally distinct triblock terpolymers into the casting dope
solution. In this way, the ability to dial in different chem-
istries and their associated functionalities to the pore sur-
face is accomplished during the standard fabrication step,
completely eliminating extra post-modification steps.
Combining different chemistries into the pore wall may
enable new and attractive capabilities in selectivity, e.g.,
in complex protein separations.
2. Results and Discussion
2.1. Triblock Terpolymer Characterization
For this study, two triblock terpolymers, poly(isoprene-
b-styrene-b-(4-vinyl)pyridine) (ISV) and poly(isoprene-b-
styrene-b-(dimethylamino)ethyl methacrylate) (ISA), of
similar molar mass and volume fractions, were synthe-
sized by sequential anionic polymerization. The molar
mass of ISV and ISA was 119 and 118 kg mol−1, respectively.
ISV had volume fractions of 0.22, 0.64, and 0.14, for the
polyisoprene (PI; ρ = 0.913), polystyrene (PS; ρ = 1.05), and
poly-4-vinylpyridine (P4VP; ρ = 1.15) blocks, respectively.
Similarly, ISA had volume fractions of 0.23, 0.63, and 0.14,
for the PI, PS, and poly(dimethylamino)ethyl methacrylate
(PDMAEMA; ρ = 1.18) blocks, respectively. Figure 1a,b
shows the chemical structures of ISV and ISA together with
a table summarizing the polymer characterization results.
2.2. Membrane Preparation, Fabrication,
and Characterization
Membranes were fabricated by a hybrid process of block
copolymer SNIPS. The triblock terpolymers were dissolved
in an appropriate solvent system. The solvent system
was chosen similar to previous studies on ISV terpolymer
derived asymmetric membranes.[6] In order to generate
blended membranes, ISV and ISA were separately dis-
solved in the binary solvent system of 1,4-dioxane (DOX)
and tetrahydrofuran (THF) in a 7:3 ratio (by weight)
(7:3 DOX/THF). For pure membranes, ISV was dissolved in
7:3 DOX/THF while ISA was dissolved in dimethylforma-
mide (DMF) and THF in a 7:3 ratio (by weight) (7:3 DMF/
THF) as described in Figure 1c. A change of the solvent
system for pure ISA membranes was necessary in order
to fabricate mechanically stable and cohesive membranes
enabling pH-stimulus responsive performance studies
(see Figure S1a, Supporting Information).
All membranes were cast from dope solutions with a
final polymer concentration of 15 wt%. The membrane
casting solutions contained different ISV:ISA weight
ratios: 1:0, 9:1, 7:3, 6:4, and 0:1. Majority ISA blended
membranes could not be studied as the SNIPS process did
not lead to mechanically stable structures (see Figure S1b,
Supporting Information). For casting solutions containing
both triblock terpolymers, final dope solutions were
prepared by mixing ISV and ISA polymer solutions and
allowed to stir at 200 rpm for 10 min before casting.
The 15 wt% dope solutions were cast onto a glass sub-
strate using an automated blade-casting machine. The
thin films were evaporated for 100 s. This evaporation
period creates a concentration gradient in the film normal
direction driving the self-assembly of block copolymers
near the top surface to produce the selective skin layer
while the bottom structure remains disordered resulting
in a sponge-like substructure providing mechanical sta-
bility upon plunging into the water precipitation bath.
In order to correlate the addition of ISA, relative to ISV,
to the casting solution with membrane structure and per-
formance, three ISV:ISA mixing ratios (by weight) were
employed: 9:1, 7:3, and 6:4. Pure ISV and ISA membranes
were fabricated to serve as references. A schematic of pure
and blended membrane top surface structures is shown
in Figure 1d where the different terpolymers are depicted
as spheres of different color (blue: ISV; green: ISA) and
Macromol. Rapid Commun. 2016, 37, 1689−1693

3. Macromolecular
Rapid CommunicationsAsymmetric Membranes from Two Chemically Distinct Triblock Terpolymers. . .
www.mrc-journal.de
www.MaterialsViews.com 1691© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
for simplicity only two layers of terpolymer micelles
are depicted. This schematic reflects the cubic block
copolymer micelle morphology that is at the origin of
the square pore lattice observed for ISV based BCP UF
membranes.[6,17,18]
Scanning electron microscopy (SEM) images of the
resultant membranes’ top surface (top row), cross section
(middle row), and areas close to the top surface (45° tilted
image, bottom row) are shown in Figure 2. All five mem-
branes are composed of a porous top surface spanning
the skin layer thickness (≈200 nm) above a hierarchically
porous, sponge-like substructure. The pure ISV membrane
(Figure 2a) exhibited a high density of uniform pores
arranged in a 2D square lattice typically associated with
ISV membranes (see Figure S2, Supporting Information).[16]
The surface structure reflects the underlying cubic block
copolymer microphase separated lattice as demonstrated
by Gu et al.[18]
With the introduction of ISA (10% and 30%;
Figure 2b,c), the top surface structure remains ordered.
Both blended membranes retain a high density of well-
ordered and packed pores. For the 5:5 blend a marked loss
of surface order is already detected in parts of the mem-
brane surface consistent with a disordered surface struc-
ture for pure ISA membranes (Figure 2d,e).
2.3. Membrane Performance
In order to correlate composition with membrane perfor-
mance, pH dependent permeabilities were measured for
pure and blended membranes using a pressurized dead-
end stirred cell (see Supporting Information, Methods
section). Normalized permeability results from the flow
experiments are presented in Figure 3 (for absolute per-
meability values see Figure S3, Supporting Information).
In a previous study it was shown that the permeability
of ISV membranes is a strong function of pH due to pH
dependent protonation and chain stretching of the P4VP
brushes at the pore surface.[6]
This is consistent with the
pH-responsive permeability observed for the pure ISV
membrane in Figure 3.
Above the pKa of 4.6 of P4VP,[6]
the permeability is
high due to the low degree of protonation resulting in
the collapse of the P4VP chains. Around the pKa of P4VP,
the permeability strongly decreases due to the increased
protonation of P4VP causing the chains to extend toward
the center of the pore and therefore, hindering transport
through the membrane.[19]
A similar behavior is observed
in Figure 3 for ISA membranes where PDMAEMA chains
reside at the pore walls. However, since PDMAEMA has a
pKa of 7.8,[3]
the curve is shifted to higher pH values rela-
tive to behavior of ISV.
Blended membranes exhibit a mixed performance
intermediate between that of pure ISV and ISA mem-
branes. In particular, a pronounced shift to higher pH
values of the pH dependent behavior is observed in mem-
branes with increasing amounts of ISA. For example, the
onset of permeability reduction occurs at a higher pH
for 9:1, and at even higher pH for 7:3 and 6:4 blended
Macromol. Rapid Commun. 2016, 37, 1689−1693
Figure 1. Pure and blended mesoporous membranes are derived from poly(isoprene-b-styrene-b-(4-vinyl)pyridine) (ISV) and poly(isoprene-
b-styrene-b-(dimethylamino)ethyl methacrylate) (ISA) triblock terpolymers. a) Chemical structures of ISV and ISA. b) Table displaying molar
mass, volume fractions (f), and polydispersity index (PDI) of the two terpolymers. c) Terpolymers and their respective solvent systems.
d) Schematic of formation of pure and blended membranes casted by block copolymer self-assembly and nonsolvent induced phase
separation (SNIPS) process. For simplicity, ISV and ISA are indicated as blue and green spheres, respectively, to visualize the compositional
variations within the membrane. The true distribution of the terpolymers in the membrane is currently unknown.

4. Y. M. Li et al.
Macromolecular
Rapid Communications
www.mrc-journal.de
www.MaterialsViews.com1692 © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
membranes relative to pure ISV membrane. Details of
the changes of the permeability curves relative to each
other are also instructive. In all blended membranes, the
majority polymer is ISV. This is consistent with the obser-
vation of decreasing low pH sections/steps in the curves
that resemble ISV behavior as a function of increasing
ISA weight fraction. In contrast, the high pH sections of
the blended membrane permeability curves do not show
pure ISA behavior but rather shift to lower pH values as
a whole (vide supra). When normalized permeability for
the blended membranes is plotted versus wt% ISA for a
pH value around 6, i.e., midway between the pKa values of
ISV and ISA (see inset Figure 3), the sensitivity to composi-
tion becomes quite apparent. Finally, at pH 3, the full pro-
tonation of PDMAEMA and P4VP segments are achieved
enabling the full extension of all charged chains resulting
in minimum permeability. The same trends and behav-
iors were observed in an independent replicate of mem-
brane formation and measurements for pure ISV and ISA
and blended (9:1, 7:3, and 6:4 blends of ISV and ISA) mem-
branes (see Figure S4, Supporting Information).
3. Conclusions
In summary, SNIPS was utilized to fabricate blended
membranes comprised of two chemically distinct triblock
terpolymers, ISV and ISA. The asymmetric blended mem-
branes possess an ordered surface structure packed in a 2D
square pore lattice above a hierarchically porous sponge-
like substructure. Their pH dependent behavior is based
Macromol. Rapid Commun. 2016, 37, 1689−1693
Figure 3. Permeability (normalized) for pure (ISV,ISA) and blended
(9:1, 7:3, and 6:4 blends of ISV + ISA) membranes at various pH
values. Indicated errors are standard deviations from three repli-
cate measurements. Inset: Effect of composition on permeability
at pH = 6.The associated curve is used only as a guide for the eye.
Figure 2. SEM characterization of surface structures (top row), cross sections (middle row), and areas close to the surface (45° tilted images)
of a,f,k) 15% ISV in 7:3 DOX/THF, b,g,l) 15% 9:1 blend in 7:3 DOX/THF, c,h,m) 15% 7:3 blend in 7:3 DOX/THF, d,i,n) 15% 6:4 blend in 7:3 DOX/THF,
and e,j,o) 15% ISA in 7:3 DMF/THF derived membranes.

5. Macromolecular
Rapid CommunicationsAsymmetric Membranes from Two Chemically Distinct Triblock Terpolymers. . .
www.mrc-journal.de
www.MaterialsViews.com 1693© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Macromol. Rapid Commun. 2016, 37, 1689−1693
on the relative percentage of ISV to ISA, enabling the tai-
loring of transport properties and tunability in the gating
mechanism.
The generation of blended membranes by simply
“mixing and matching” two chemically distinct tri-
block terpolymers in the casting solution demonstrates
a pathway to advanced asymmetric block copolymer
derived UF membranes in which different pore surface
chemistries and associated functionalities can be inte-
grated into a single membrane via standard membrane
fabrication, i.e., without requiring laborious post-fabrica-
tion modification steps. As the present proof of principle
experiments have shown these mixed chemistries result
in performance profiles different from those of mem-
branes obtained from either of the two constituting block
copolymers alone. The work poses a number of scientific
questions for future studies concerning, e.g., the limits
of this blending approach in terms of allowable differ-
ences in molar mass and block fractions. It may also have
interesting implications for separation applications. For
example, blending two distinct chemistries into the mem-
brane surface may provide affinities to proteins that are
different to those of membranes obtained from either of
the two constituting block copolymers alone, and there-
fore may lead to improved selectivity. And the blending
may not be confined to only two block copolymers but
may be extended to multiple (i.e., more than two) chemi-
cally distinct block copolymers and block copolymers
with different end functionalities,[8]
enabling to tailor
membrane properties in unprecedented ways.
Supporting Information
Supporting Information is available from the Wiley Online
Library or from the author.
Acknowledgements: This project was funded by Contract No.
HDTRA1-13-C0003 from the Defense Threat Reduction Agency
(DTRA). Y.G. was funded by the National Science Foundation
(DMR-1409105). This work utilized the Cornell Center for
Materials Research (CCMR) facilities supported by NSF MRSEC
program (DMR-1120296).
Received: July 13, 2016; Revised: July 22, 2016;
Published online: September 8, 2016; DOI: 10.1002/marc.201600440
Keywords: asymmetric membranes; blending; block copolymers;
self-assembly; SNIPS
[1] K. V. Peinemann, V. Abetz, P. F. W. Simon, Nat. Mater. 2007,
6, 992.
[2] R. M. Dorin, D. S. Marques, H. Sai, U. Vainio, W. A. Phillip,
K. V. Peinemann, S. P. Nunes, U. Wiesner, ACS Macro Lett.
2012, 1, 614.
[3] F. Schacher, T. Rudolph, F. Wieberger, M. Ulbricht,
A. H. Muller, ACS Appl. Mater. Interfaces 2009, 1, 1492.
[4] J. Hahn, V. Filiz, S. Rangou, J. Clodt, A. Jung, K. Buhr, C. Abetz,
V. Abetz, J. Polym. Sci., Part B: Polym. Phys. 2013, 51, 281.
[5] M. Karunakaran, R. Shevate, K. V. Peinemann, RSC Adv. 2016,
6, 29064.
[6] W. A. Phillip, R. M. Dorin, J. Werner, E. M. Hoek, U. Wiesner,
M. Elimelech, Nano Lett. 2011, 11, 2892.
[7] J. L. Weidman, R. A. Mulvenna, B. W. Boudouris, W. A. Phillip,
J. Am. Chem. Soc. 2016, 138, 7030.
[8] Q. Zhang, Y. Gu, Y. M. Li, P. A. Beaucage, T. Kao, U. Wiesner,
Chem. Mater. 2016, 28, 3870.
[9] P. Madhavan, K. V. Peinemann, S. P. Nunes, ACS Appl. Mater.
Interfaces 2013, 5, 7152.
[10] S. P. Nunes, A. R. Behzad, B. Hooghan, R. Sougrat,
M. Karunakaran, N. Pradeep, U. Vainio, K. V. Peinemann, ACS
Nano 2011, 5, 3516.
[11] H. Yu, X. Qiu, N. Moreno, Z. Ma, V. M. Calo, S. P. Nunes,
K. V. Peinemann, Angew. Chem., Int. Ed. 2015, 54, 13937.
[12] Y. Gu, U. Wiesner, Macromolecules 2015, 48, 6153.
[13] M. Radjabian, V. Abetz, Adv. Mater. 2015, 27, 352.
[14] Y. Gu, R. M. Dorin, U. Wiesner, Nano Lett. 2013, 13, 5323.
[15] S. A. Hesse, J. G. Werner, U. Wiesner, ACS Macro Lett. 2015, 4,
477.
[16] J. I. Clodt, V. Filiz, S. Rangou, K. Buhr, C. Abetz, D. Höche,
J. Hahn, A. Jung, V. Abetz, Adv. Funct. Mater. 2013, 23, 731.
[17] R. M. Dorin, W. A. Phillip, H. Sai, J. Werner, M. Elimelech,
U. Wiesner, Polymer 2014, 55, 347.
[18] Y. Gu, R. M. Dorin, K. W. Tan, D. M. Smilgies, U. Wiesner,
Macromolecules 2016, 49, 4195.
[19] S. Qu, T. Dilenschneider, W. A. Phillip, ACS Appl. Mater. Inter-
faces 2015, 7, 19746.